NAD
ANALOGS
AND
GLYCERALDEHYDE PHOSPHATE DEHYDROGENASE
Hanson, K . R. (1966), J . Amer. Chem. SOC.88,2731. Jautelat, M., Grutzner, J. B., and Roberts, J. D. (1970), Proc. Nut. Acad. Sci. U. S. 65,288. Johnson, R . A., Hardman, J. G . , Broadus, A. E., and Sutherland, E. W . (1970), Anal. Biochem. 35,91. Kenyon, G. L., and Westheimer, F. H. (1966), J . Anzer. Chem. SOC.88,3557. Krongelb, M., Smith, T. A., and Abeles, R. H . (1968), Biochim. Biophys. Acta 167,473. Kuhn, R., and Drury, K . (1949), Ann. 564,32. Martin, D. J., Gordon, M., and Griffin, C. E. (1967), Tetrahedron 23,1831. McElroy, W. D. (1963), Methods Enzymol. 6,445. Meister, A. (1950), J . B i d . Chem. 184,117. Mildvan, A. S., and Leigh, R . A. (1964), Biochim. Biophys. Acta 89,393. Moureu, H., Chovin, P., Garein, M., and Ventrillard, J. (1952), Bull. SOC.Chim. France, 296. Nowak, T., and Mildvan, A. S. (19701, J . Biol. Chem. 245, 6057. Randerath, K . , and Randerath, E. (1964), J. Chromatog. 16,111.
Reynard, A. M., Hass, L. F., Jacobsen, D. D., and Boyer, P. D. (1961), J . Biol. Chem. 236,2277. Rose, I. A. (1960), J . Biol. Chem. 235,1170. Rose, I. A. (1970), J . B i d . Cheni. 245,6052. Rose, I. A., O’Connell, E. L., Noce, P., Utter, M. F., Wood, H. G . , Willard, J. M., Cooper, T. G., and Benjiman, M. (1969), J . B i d . Chem. 244,6130. Sprinson, D. B., and Chargaff, E. (1946), J. Biol. Chern. 164, 417. Sternlicht, H., Kenyon, G. L., Packer, E. I.,., and Sinclair, J. (1971), J . Amer. Chem. SOC.93,199. Stiles, A. R., Reilley, C. A., Pollard, G . R., Tieman, C. H., Ward, L. F., Phillips, D. D., Soloway, S. B., and Whetstone, R . R . (1961)) J . Org. Chem. 26,3960. Supniewski, J. V., and Salzberg, P. L. (1941), in Organic Syntheses, Gilman, H., and Blatt, A. H., Ed., Collect. Vol. I, New York, N. Y., Wiley, p 46. Tietz, A., and Ochoa, S. (1959), Arch. Biochem. Biophys. 78, 477. Varner, J. E. (1957), Methods Enzymol. 3, 397. Woods, A. E., O’Bryan, J. M., Mu, P. T. K., and Crowder, R . D. (1970), Biochemistry 9,2334.
Interaction of Nicotinamide-Adenine Dinucleotide and Its Analogs with Glyceraldehyde 3-Phosphate Dehydrogenase* Denise Ebyt and M. E. Kirtleyz
ABSTRACT: Several structural analogs of N A D were tested for their effects on the properties of rabbit muscle glyceraldehyde 3-phosphate dehydrogenase. Of the 13 analogs tested only three were effective substrates: acetylpyridine-AD, thionicotinamide-AD, and deamino-NAD. Several other analogs including ADP-ribose, ADP, AMP, N M N , and pyridine carbaldehyde-AD were inhibitors competitive with NAD. The latter analog was also competitive with the substrate phosphoglyceraldehyde. Two of the analogs gave a spectrum with the enzyme similar to that observed with p-
I
n recent years there has been interest in the unusual kinetic and binding properties of rabbit muscle glyceraldehyde 3phosphate dehydrogenase (GPDH)’ and the corresponding
* From the Department of Biological Chemistry, University of Maryland, School of Medicine, Baltimore, Maryland 21201. Receiced December 28, 1970. This investigation was supported by Grant A M 11428 from the National Institutes of Health and by Grant GB-8362 from the National Science Foundation. t Present address: St. Joseph’s College, Emmitsburg, Md. 21727. t T o whom to address correspondence. Abbreviations used are: GPDH, glyceraldehyde 3-phosphate dehydrogenase (EC l .2.1.12); ADP-ribose, adenosine diphosphoribose; acetylpyridine-AD, acetylpyridine-adenine dinucleotide; thionicotinamide-AD, thionicotinamide-adenine dinucleotide; deamino-NAD, nicotinamide-hypoxanthine dinucleotide; deaminoacetylpyridine-AD, acetylpyridine-hypoxanthine dinucleotide; DTT, dithiothreitol; IAA, iodoacetic acid.
NAD. These compounds, acetylpyridine-AD and deaminoNAD, also showed anticooperative binding to the enzyme. Several analogs produced partial stabilization of the enzyme at 45’. I n decreasing order of stabilization these are: @-NAD, ADP-ribose, thionicotinamide-AD, ADP, NMN. From these studies as well as determinations of the kinetic and binding constants of the analogs it is possible to deduce a probable conformation of the bound coenzyme and to specify the major points of contact between the coenzyme and the enzyme. A model of the enzyme-coenzyme complex is presented.
enzyme obtained from yeast. Kirschner et a/. (1966) have reported that the yeast G P D H exhibits positive cooperative binding of the coenzyme N A D under some conditions while the muscle enzyme does not give cooperative binding. On the other hand, Conway and Koshland (1968) have shown that there is anticooperative binding of N A D to the muscle enzyme. In a more recent study Cook and Koshland (1970) have reexamined the binding of N A D t o the yeast enzyme and have shown that there is positive cooperativity in the first half of the curve and negative or no cooperativity in the second half of the curve. These obvious functional differences between the yeast and muscle enzymes are particularly interesting because of the high degree of homology in the primary structures of GPDHs from different species (Harris and Perham, 1968). One possible explanation for the functional differences is that the N A D BIOCHEMISTRY,
VOL.
IO,
NO.
14, 1 9 7 1
2677
EBY AND
binding site of the yeast enzyme might interact with the coenzyme quite differently than does the binding site of the muscle enzyme. In order to test this possibility we have examined the binding of several coenzyme analogs t o muscle G P D ; Stockell (1959) has previously reported similar studies for the yeast enzyme. The data reported in this paper show that the direrences in binding of the analogs to the two enzymes arc small and cannot account for the functional dissimilarity of these enzymes. From the analog studies with G P D H together with the recent reports on the preferred conformations of N A D and its analogs (Sarma and Kaplan, 1968, 1970a,b; Sarma et ul., 1968, 1970). it has been possible to draw some conclusions about the contact points between G P D H and N A D and about the probable conformation of N A D when it is bound to the enzyme. This structure is discussed and compared to the known structure of the lactate dehydrogenase-NAD complex (Adams e t a l . , 1969, 1970). Experimental Section Materials. Crystallized rabbit muscle glyceraldehyde 3phosphate dehydrogenase was purchased as an ammonium sulfate suspension from Sigma Chemical Co. Glyceraldehyde ?-phosphate was obtained as the DL-glyceraldehyde 3-phosphate diethyl acetal barium salt from Sigma. The compound was deionized on Dowex 50-X4 and the resulting solution heated to 100' for 3 min t o obtain the free aldehyde. @-NAD and N A D analogs were also from Sigma. N A D analogs included: a-NAD, NADP, AMP, ADP, ATP, N M N , ADPribose, acetylpyridine-AD, thionicotinamide-AD, deaminoNAD, deamino-NADP, deaminoacetylpyridine-AD. Dithiothreitol was from Calbiochem. All other reagents were analytical grade. Cliurcoul Treutriietir oj' GPDH. A solution of the enzyme (about 3 mg/ml) in 0.01 SI phosphate-1 mhi EDTA-1 mhi DTT (pH 7.0) was treated with 3-4 mg of activated charcoal1 mg of protein (Velick, 1953). This produced apoenzyme with A2skll!A2,jo of 1.8-1.9 and with about 90% of the activity of untreated enzyme. Based o n the results of Fox and Dandliker ( 1 965) this absorbance ratio represents 0.1 5-0.2 equiv of N A D bound per equiv of enzyme. The molecular weight of G P D H was taken to be 140,000 (Harris and Perham, 1965). The protein concentration of apoenzyme and holoenzyme was determined using the 280-mp extinction coefficients of Murdock and Koeppe (1964), 0.815 and 1.06 cm2 mg-I for apoand holoenzymes, respectively. E ~ i z j , / ~ u tAssuj.s. ic G P D H activity was measured with a Zeiss P M Q I1 spectrophotometer according to a modification of the standard method (Velick, 1955). The assay mixture contained 2.0 nihi @-NAD,2.0 i n M DL-glyceraldehyde 3-phosphate, 10 ma! sodium arsenate, 10 m.hi EDTA, and 0.066 M sodium pyrophosphate with a final p H of 8.0 in a total volume of 1.0 ml. Assays were initiated by addition of 10 pl of the enzyme and the increase in absorbance at 340 mH was followed at 10-sec intervals t o obtain the initial rate. The specific activity of G P D H was expressed as micromoles per minute per milligram of protein. Coenzjxie Analog Substrate Acticil>,. The standard enzyme assay was used except that the assay mixture was made 2 mhi with respect to the N A D analog rather than 8-NAD. Substrate activity was expressed as per cent activity in the reduction reaction relative to /3-NAD activity at the same concentration. K,,, and V,,,,,,were determined from double-reciprocal plots.
2678
BIOCHEMISTRY,
VOL.
IO,
NO.
14, 1 9 7 1
KIRTLEY
Znhibiror Constunts. Inhibition studies were carried out under the conditions of the standard assays but with N A D concentrations in the range of 4 X to 4 X IO-' hf or glyceraldehyde 3-phosphate concentrations in the range of 7 X 10-j to 7 X lo--' hi and inhibitor concentrations above and below the Ki values. In all cases at least five concentrations of substrate and four concentrations of inhibitor were used. Lineweaver-Burk and Dixon plots from experiments that employed varying substrate concentrations performed at several different inhibitor concentrations were used to determine K , values. Equilibrium Diulysis. Dialyses were carried out in 1 .O-ml cells for 24 hr at 1' with gentle shaking. Controls showed that equilibrium was reached in that time. Ligand concentration on both sides of the membrane was determined from the extinction coefficient at the maximum absorbance wavelength for the analog. About 1.8 mgiml of apoenzyme was dialyzed against N A D analog in 0.01 hi phosphate-I miv EDTA-1 mhi DTT (pH 7.0) buffer. The , ,A, and corresponding extinction coeficient for the analogs used are: @-NAD, 17,800 at 259 mp;thinicotinamide-AD, 19,700 at 259 mp (Seigel et d., 1959; Anderson erul.. 1963). Sl)ecrroliliototiier~ir Titrution. Dialyzed solutions of apoenzyme (1.0--2.0 mg,'ml) in 0.01 M phosphate--1 mM EDTA-I i n M D T T (pH 7.0) were scanned from 500 to 300 m p in a Cary 14 spectrophotometer using the 0-0.1 slide-wire. Absorbance a t 360 mp was measured as a function of added N A D analog, and absorbance of the free analog was compensated for by adding equal aliyuots of analog to the reference cuvet. The concentration of analog required to saturate the total GPDH sites was determined graphically from a plot of absorbance a t 360 mp cs. concentration of added analog. The per cent saturation at lower analog concentrations was calculated as the ratio of the 360-mp absorbance at thtse levels to the absorbance at saturation. Tlierm~lStubiliij,. The thermal stability of charcoal-treated G P D H was determined by incubating 1.0 ml of the enzyme (0.2-0.3 mg/ml) in 0.01 M phosphate-1 nihi DTT-1 mki EDTA (pH 7.0) at 45" and periodically assaying 10-p1 aliquots for enzyme activit) . The first-order inactivation rate constants were obtained from a semilog plot of per cent initial activity cs. time of incubation in the presence or absence of I00 phi analogs. Each GPDH- analog rate constant was compared with that obtained simultaneously for the pure apoenzyme prepared for the particular experiment and the results were expressed as rate constant ratios klnnalog)/k(no analog).
Results Substrute Acricity. Of the 13 analogs tested, only 3 were effective as substrates in the oxidation of phosphoglyceraldehyde under standard assay conditions : acetylpyridine-AD, thionicotinamide-AD, and deamino-NAD. The values for K,,, and V,,,,, for these analogs are shown in Table I . The three analogs which are effective substrates gave normal Michaelis-. Menten kinetics and showed n o evidence of cooperative or anticooperative behavior under the conditions used in these assays. Two other analogs, acetylpyridinedeamino-ADand NADP, showed slight activity with the enzyme. At a concentration of 2 mha both gave 1 of the activity of N A D when assayed with the apoenzyme but there was no activity with the holoenzyme. Znhibifion. Several other analogs are inhibitors of the reaction and in all cases these are competitive with NAD. The K ; values are given in Table 11. In the group of analogs which
NAD ANALOGS AND GLYCERALDEHYDE PHOSPHATE DEHYDROGENASE ~
o.o+
Kinetic and Binding Constants for NAD Analogs as Substrates for GPDH: TABLE I:
1-
0,03
Analog
K, ( p M )
@-NAD Acetylpyridine- AD Thionicotinamide- AD Deamino-NAD
50 500 310 2300
Re1
Vmax
Kd
1. o 0.1 0.2 0.8
(W) 16
98 112
a The V,,, values are the ratios of V,(analog) : V,(P-NAD). The following analogs gave no activity : pyridinecarbaldehyde-AD, NMN, N M N AMP, a-NAD, deamino-NADP. The activity observed with N A D P and acetylpyridinedeaminoNAD was too low and required too high concentrations to permit accurate determinations of Km and V,n,,.
,100
+
TABLE 11:
Inhibitor Constants for N A D Analogs:
200
300 ANALOG
400
600
500
pm
1 : Titration of charcoal-treated rabbit muscle GPDH with NAD analogs. Charcoal-treated GPDH, A I s a : A e s a= 1.80, was titrated to maximum absorbance at 360 mp. Enzyme concentration was 14 /IMwith a specific activity of 28. Enzyme and analogs were dissolved in 0.01 M phosphate-1 mM EDTA--1 mM DTT (pH 7.6).
FIGURE
76-
0.025 0.18 2.5 2.5 2.7 2.8 2.9
Pyridinecarbaldehyde-AD ADP-ribose AMP AMP NMN NADP NMN ADP
+
a The other analogs which were not active as substrates (a-NAD, deamino-NADP, and acetylpyridinedeamino-NAD) were not inhibitors.
represent partial structures of N A D the most effective inhibitor is ADP-ribose. The mixture of A M P N M N contains all the moieties of NAD except the pyrophosphate bond and this mixture is no more inhibitory than AMP alone. By far the most effective inhibitor of all the analogs tested is pyridinecarbaldehyde-AD. This analog is unusual in that it is competitive not only with NAD but also with the substrate phosphoglyceraldehyde. NADP is a weak inhibitor competitive with NAD. All other analogs which were not substrates gave no inhibition when tested at concentrations up to 4 mM. Equilibrium Dialysis Studies.The values for the dissociation constants for @-NAD, thionicotinamide-AD, and deaminoNAD as measured by equilibrium dialysis are given in Table I. In order to reach equilibrium it was necessary to carry out the experiments for a t least 18 hr. @-NAD, thionicotinamide-AD, and deamino-NAD were stable over this period but it was not possible to use acetylpyridine-AD in these experiments. Over a period of several hours at either 25 or 4' aqueous solutions of this analog developed a yellow color and gave a progressive increase in absorbance a t 260 mp. The value obtained for pNAD is comparable to that reported by Velick (1958) and the value of K8 obtained by Conway and Koshland (1968)' The K d for thionicotinamide-AD a t 4 " is about one-third that of its K,. Deamino-NAD is quite different from p-NAD or
5-
4-
321-
I
!
t
1.0
2.0
-V
I
I
3 .O
4.0
FIGURE2 : Scatchard plot
for the binding of deamino-NAD t o rabbit muscle GPDH. Four independent and equivalent binding sites for the analog were assumed in the calculation but the curvature of the plot indicates that there is anticooperativity in the binding of this analog to multiple sites.
+
2 The conditions used by us in this experiment and by Velick were such that only K1 is measured.
thionicotinamide A D in that its K,,, is at least 20 times higher than the Kd measured by equilibrium dialysis. Spectrophotometric Properties of the Enzyme Analog Coniplexes. Racker and Krimsky (1952) first showed that the binding of P-NAD to the enzyme produces a broad absorbance band with a maximum at 360 mp. Figure 1 shows the increase in absorbance at 360 mk produced by successive additions of the two analogs which gave this effect. From these data it is possible to obtain values of per cent saturation as a function of the concentration of added analog. Although there are known to be four binding sites for N A D and one can assume that there are also four binding sites for the analogs, these data do not define how many of the actual binding sites contribute to the absorbance at 360 mh. When the data for deamino-NAD were calculated on the basis of four binding sites and the data plotted on the Scatchard plot the curve shown in Figure 2 was obtained. The marked curvature of the plol indicates that if all four sites are contributing to the absorbance the binding constants are not identical. The corresponding data obtained for acetylpyridineN A D are shown in Figure 3. When the same data were recalculated assuming one, two, or three effective binding sites, similar sharp curvatures in the Scatchard plots were obtained. Treatment of the data for four sites by the method of Conway and Koshland (1968) to obtain the dissociation constant for BIOCHEMISTRY,
VOL.
10,
NO.
14, 1 9 7 1
2679
EBY AND
Effect of N A D Analogs o n Thermal Inactivation of GPDH..
TABLE III:
I
IL
Analog /3-NAD ADP-ribose Thionicotinamide-A D ADP NMN AMP NMN
lr
#b
zo
KIRTLEY
-
Jio
7k-
3 : Scatchard plot for the binding of acetylpyridine-AD to rabbit muscle GPDH.
FIGURE
+
Rate Constant Ratio X(analog)/k(no analog) 0 0 17
0.25 0.50 0 50 0.50
The concentration of apoenzyme was 0.3 mg/ml and all analogs were 0.1 m x The analogs which failed to stabilize the enzyme were : deamino-NAD, acetylpyridine-AD, pyridinecarbaldehyde-AD, AMP, NADP, a-NAD, deaminoNADP, and acetylpyridinedeamino-AD. 0
the final site failed to produce a linear Scatchard plot. Because of the uncertainty of the number of binding sites which produce the measured spectral change one can say with certainty only that there is more than one of these and they are not independent and identical. If one assumes four binding sites and calculates the average K,I between 50 and 80% saturation, a value of about 40 p~ is obtained and the site or sites a t lower per cent saturation must have K,, a t least five times lower than this. Tliernial Stabilitj~Yf the Enzj~r?ie-Anulog Complexes. The apoenzyme of GPDH is markedly unstable at neutral pH both at 4 ” and at elevated temperatures while the enzyme with bound N A D is stable under these conditions (Grisolia and Joyce, 1959). Each of the analogs was tested to determine whether it affected the thermal stability of the apoenzyme. It was found that incubation of the apoenzyme under the conditions shown in Figure 4 gave a first-order decay in enzyme activity. Each analog was tested a t a concentration of 100 p~ for its efTect on the stability of the enzyme and the loss of activity was determined as a function of time. When appropriate, other concentrations of the analogs were employed as well. In all cases the loss in enzyme activity followed first-order kinetics and values of the rate constants were obtained. The reported values in Table 111 are the ratios of these rate constants in the presence and absence of the analog. Under the conditions used there was no loss of enzyme activity when the enzyme was saturated with $-NAD and consequently the ratio for P-NAD
IS
30 4 5
60 75 90 105 120 I35 IS0 I65 MINUTES
4: Effect of NAD and NAD analogs on the stability of GPDH at 45’. Protein concentration is 0.3 mg/ml in 0.01 hi phosphate-1 mM EDTA (pH 7.0). The charcoal-treated protein was incubated at 45‘ and aliquots were removed for assay under the standard conditions at the indicated times. The additions were: ( 0 ) none, (0) 100 pihi NAD. (A) 100 phi ADP-ribose. (A) 100 p~ thionicotinamide-AD, ( X ) 100 +I NMN, and ( 2 )100 p i ADP.
is zero. For analogs which give no protection against inactivation the ratio is 1.O. The analogs most etrective in stabilizing the enzyme were ADP-ribose and thionicotinamide-AD. A D P afforded partial protection as did N M N . 3 A M P did not stabilize the enzyme and the mixture N M N + A M P was no more effective than N M N alone. All other analogs tested (Table 111) failed to stabilize the enzyme even when used a t concentrations sixfold higher than those reported here. Discussion Compurisons to Yerist GPDH. Although the G P D H from yeast and muscle are similar in many respects, it has been shown that they differ in one important function, namely, the yeast enzyme exhibits cooperative binding of NAD while the muscle enzyme shows anticooperative binding of the coenzyme (Cook and Koshland, 1970; Kirschner ef al., 1966; Conway and Koshland, 1968). It was thought that a comparison of the two enzymes using N A D analogs might provide some insight into this functional difference; however the date summarized in Table IV shows that the two enzymes are very similar a t least with respect to the three parameters for which comparable data are available. Both enzymes produce a n absorbance band at 360 mp when complexed with @-NAD or with two analogs which are effective substrates. The values of the dissociation constants for the analogs and 8-NAD are very similar for the two enzymes. It should be emphasized that these values, as obtained by equilibrium dialysis or inhibition studies (muscle enzyme) or by spectrophotometric titration (yeast enzyme), are the binding constants for the final binding site of both enzymes. Because of the anticooperative binding of the muscle enzyme the KdlssOcn for the first binding site is several orders of magnitude lower than is that of the yeast enzyme. I n the spectrophotometric titration studies carried out with the muscle enzyme we have shown that there is anticooperative binding of the two analogs which are effective substrates but that the
FIGURE
2680
BIOCHEMISTRY,
VOL.
10,
NO.
14, 1 9 7 1
a Although it is clear that XMN binds to the muscle enzqme thcri: is n report that N M N does not inhibit the yeast GPDH (Yang a n d Deal, 1969). This difference is not reflected in the results with any of the other analogs and remains an anomaly.
NAD ANALOGS AND
TABLE I V : Comparison
GLYCERALDEHYDE PHOSPHATE DEHYDROGENASE
of the NAD Analogs to Yeast and Muscle GPDH: Absorbance of Complex at 360 mp
Re1 V,,,
KdiSSOCIl
....
Analog
Muscle
Yeast
Muscle
Yeast
Muscle
Yeast
P-NAD Acetylpyridine-AD Deamino-NAD Pyridinecarbaldeh yde-AD Acetylpyridinedeamino-AD Thionicotinamide-AD
Yes Yes Yes No No
Yes Yes Yes No No
16 p~ 500 112 25
7 . 8 p~ 260 130 13 4800
100 10 80 0 1 20
100 8.7 8.3 0 1.7
98
= The data for the yeast enzyme are those reported by Stockell (1959).
binding to the final site has a Kdissocnsimilar to that of the yeast enzyme. The only obvious discrepancy between the results with the yeast and the muscle enzymes is the v,,, value for deamino-NAD. In all other respects this analog gives similar results with the two enzymes. From these results it appears that there is no significant qualitative difference between the coenzyme binding sites for the yeast and muscle enzymes sufficient to account for the functional difference between cooperative and anticooperative binding in the two cases. Thus it appears that the cooperative or anticooperative binding phenomena are related to differences in the structure or conformation of the enzymes outside the immediate binding sites for the coenzymes. Important Interaction Sites between N A D and Rabbit Muscle GPDH. Another result of these studies in the finding that there is a n important interaction of the ribose phosphate moiety of the N M N half of the coenzyme with the N A D binding site. For example, the dissociation constant for ADP ribose is a t least ten times lower than that of ADP or AMP. The direct interaction of the phosphate of the N M N half of the coenzyme is suggested by the fact that both N M N and A D P produce some stabilization of the enzyme while AMP does not. This stabilization effect is in part synergistic with the binding of both the adjacent covalently bound phosphate on one side and the covalently bound ribose on the other. When both adjacent residues are present, as in ADP-ribose, the resulting complex has stability approaching that of the @-NADcomplex and greatly in excess of the complexes of any one of the smaller analogs or of the mixture of N M N AMP. Previous reports have concluded that the principal form of P-NAD in aqueous solution a t neutral pH has the aromatic rings stacked (Sarma and Kaplan, 1968, 1970a; Sarma et a[.? 1970) and that the reduced coenzyme when bound to G P D H is in a conformation with the aromatic rings stacked (Velick, 1958, 1961). Although these points are still open to some criticism (cy. Jacobus, 1970), the data obtained with the N A D analogs are consistent with the binding of @-NADin a stacked conformation with a strong interaction with the enzyme at the ribose-phosphate moiety of the N M N half of the coenzyme. Of the three analogs which are effective substrates, only thionicotinamide-AD partially stabilizes the enzyme. According t o Sarma and Kaplan (1968, 1970a) acetylpyridine-AD exists in a stacked conformation, but, unlike @-NAD, the acetylpyridine ring is in the anti rather than the syn conformation with resulting changes in the conformation of the PPR backbone. Deamino-NAD occurs primarily in the open conformation (Sarma and Kaplan, 1968, 1970a). The unusually
+
high K, for this analog and its failure to stabilize the enzyme suggest that only the stacked conformation is active as a substrate. In trying to obtain an approximate picture of the NAD binding site we have used the molecular model of the stacked form of the coenzyme in the P helix. The model shown schemactically in Figure 5 is consistent with all data which are presently available. It is notable that in this orientation of the P helix the substituents of the ribose ring attached to nicotinamide moiety project back into the proposed site and the ionizable groups of the pyrophosphate backbone project to the left and back of the site. Comparisons to the Coenzjwe Binding Site of Lactate Dehydrogenase. There are some clear distinctions which can be made between the coenzyme binding site of G P D H and that of lactate dehydrogenase. The binding of AMP to lactate dehydrogenase is a prerequisite to the binding of N M N (McPherson, 1970), in contrast with our results with GPDH. According to McPherson the space group change of crystalline lactate dehydrogenase produced by a mixture of N M N AMP differs from that produced by AMP alone and there is no change produced by N M N alone. We have observed no synergistic effects of N M N AMP on the properties of GPDH but we d o find effects of N M N or AMP alone. Velick (1958, 1961) in his fluorescence studies of lactate dehydrogenase and G P D H concluded that the coenzyme binds in an open conformation to lactate dehydrogenase and in a stacked conformation to G P D H and the crystallographic data of Adams et al. (1969, 1970) clearly confirm the first of these conclusions. It therefore appears that all data so far available are consistent with the conclusion of Velick that lactate dehydrogenase and GPDH are representatives of two different classes of dehydrogenases, one which binds NAD in an open conformation and the other which binds NAD in a stacked conformation.
+
+
4 The reason for choosing the P helix rather than the M is as follows. It has been shown that in the GPDH reaction the H is introduced from the substrate onto the “B” face of the nicotinamide ring (Vennesland and Westheimer, 1954; Levy e t al., 1962). The formation of an apparent charge transfer complex with tryptophan in the enzyme-NAD complex suggests that the exposed face of the nicotinamide ring is close to a tryptophan residue and, furthermore, there is evidence that the electron transfer in some dehydrogenases may be mediated by tryptophan (Schellenberg, 1967; Chan and Schellenberg, 1968; but see also Allison et a / . , 1969). The “B” face of the nicotinamide ring is exposed in the P helix but not the M helix of P-NAD and therefore we propose that the bound conformation of P-NAD is that of the P helix.
BIOCHEMISTRY,
VOL.
10,
NO.
14, I 9 7 1
2681
FIGURE 5 : Schematic view of p-NAD at the active site of GPDH. The enzyme is in contact with the molecule on all sides except the one facing out of the page. The pyridine ring is stacked above a tryptophan residue and in the P-helical form the amide side chain is directed toward the front of the site, A cysteine sulfhydryl group to the right of the tryptophan residue would be accessible to the substrate and the C : position of the pyridine ring as well as to various sulfhydryl reagents. The ribose ring attached to pyridine in this Phelical form is below and to the left of the pyridine ring and both OH groups project to the left and back of the molecule. Both phosphate residues are above these OH groups with one almost directly in back of the other. All the oxygen substituents of the phosphate groups except the bridge oxygen are directed either to the left or top of the molecule and, together with the ribose substituents, provide a compact region where there appears to be major contact with the site. The ribose attached to adenine projects forward and its two OH groups are directly in front of the molecule. Phosphate substituents on either of these positions would have to be positioned above the rest of the molecule and thus could hinder binding due either to steric hindrance or to electrostatic repulsion with groups on the protein. The adenine ring is stacked above the pyridine ring and in this P helix the amino group is directly behind the molecule where it would also be buried in the complex.
References Adams, M. J., Haas, D. J., Jeffery, B. A., McPherson, A , , Jr., Mermall, H. L., Rossmann, M. G . , Schevitz, R. W., and Wonacott, A. J . (1969), J . Mol. Biol. 41,159. Adams, M. J., McPherson, A., Jr., Rossmann, M. G., Schevitz, R . W., and Wonacott, A. J. (1970), J . Mol. Biol. SI, 31. Allison, W. S . , Connors, M. J., and Parker, D. J. (1969), Biocheni. Biophys. Res. Commun. 34,503. Anderson, B. M., Anderson, C. D., Lee, J. K., and Stein, A. M. (1963), Biochemistry 2, 1017. Chan, T. L., and Schellenberg, K. A. (1968), J . B i d . Chein. 243,6284. Cilento, G., and Zinner, K . (1968), in Molecular Associations in Biology, Pullman, B.. Ed., New York, N . Y . , Academic Press, p 309. Colowick, S. P., Van Eys, J.. and Park, J. H. (1966), Curnp. Bioclrem. 14. 1. Conway, A,, and Koshland, D. E., Jr. (1968), Biochemistry 7,4011.
2682
BIOCHEMISTRY,
VOL.
10,
NO.
14, 1 9 7 1
Cook, R . A,, and Koshland, D. E., Jr. (1970), Biochemistry 9, 3337. Fawcett, C. P., and Kaplan, N. 0. (1962), J . Biol. C./ie/?i. 237,1709. Fenselau, A. (1 970), J . B i d . CJiem. 245, 1239. Fox, J . B., Jr., and Dandliker, W. B. (1965), J . Biol. Cl7etn. 221,1005. Furfine, C. S.,and Velick, S. F. (19651, .I Bioi. . Chetir. 240, 844. Grisolia, S., and Joyce, B. J. (19591, Biocliem. Biopliys. Res. Commun. I , 280. Harris, J. E., and Perham, R . N. (1968), Nature (London) 219,1025. Harris, J. L., and Perham, R. N. (1965), J . Idol. Biol. 13, 186. Jacobus, J. (1971), Biochemistry IO, 161. Kaplan, N. 0.(1960), Enzymes 3,105. Kirschner, K., Eigen, M., Bittman, R., and Roight, B. (1966), Proc. Nat. Acud. Sci. U. S. 56, 1661. Kirtley, M. E., and Koshland, D. E., Jr. (1970), J . Biol. C-heni, 245,276. Levy, H . R., Talalay, P., and Vennesland, B. (1962), Progr. Stereochem. 10,299. McPherson, A , , Jr. (1970), J . Mol. Biol. 51, 39. Murdock, A. L., and Koeppe, 0. J. (1964), J . Bid. C/ie,n. 239,1983. Pfleiderer, G . , Sann, E., and Ortanderl, F. (1963), Biocliinz. Biopliys. Acta 73,39. Racker, E., and Krimsky, I. (1952), J . B i d . Cheiii. 198, 731. Sarma, R . H., and Kaplan, N. 0. (1970a), Biockernisir?, 9, 539. Sarma, R . H., and Kaplan, N. 0. (1970b), Biochemistry 9, 557. Sarma, R . H., Moore, M., and Kaplan, N. 0. (1970), Bioclieinistry 9, 549. Sarma, R . H., Ross, V., and Kaplan, N. 0. (1968), Biochemistry 7, 3052. Schellenberg, K. A . (1967),J. B i d . Chetw. 242, 1815. Shifrin, S., and Kaplan, N . 0.(1960), Adcun. Enzytnol. 22,347. Siegel, J. M., Montgomery, G . A., and Bock, R. A. (1959), Arch. Biocheni. Biopliys. 82,288. Stockell, A. (1959), J . Biol. Chem. 234,1293. Sund, H. (1968), in Biological Oxidations, Singer, T. P., Ed., New York, N. Y . ,Interscience, pp 603,641. Velick, S . F. (1953), J . Biol. C/zeni. 203, 563. Velick, S. F. (1955), Methods Enzy/nol. I , 401. Velick, S. F. (1958),J. Bid. Clieiir. 233, 1455. Velick, S. F. (1961), in Light and Life, McElroy, W. E., and Glass, B., Ed., New York, N . Y., Academic Press, p 108. Velick, S. F., Hayes, J. E., Jr., and Harting, J. (1953), J . B i d . Chem. 203, 527. Vennesland, B., and Westheher, F. H . (1954), in Mechanism of Enzyme Action, McElroy, W. E., and Glass, B., Ed., New York, N. Y., Academic Press, p 357. Yang, S. T., and Deal, W. C., Jr. (1969), Biochemistry 8, 2806.